Low-cost production of solar cells on flexible substrates using printing or web coating technologies is promising a highly cost-efficient alternative to traditional silicon-based solar cells. Recently, solar cells with absorber layers fabricated by solution-based deposition of alloys of copper (Cu) and indium (In) with selenium (Se) or sulfur (S) have been developed. Such solar cells, generally referred to as CIGS cells, have been fabricated using different non-vacuum processes in which a precursor solution is formulated containing mixed oxides of Cu, In and Ga, which is then coated on a substrate. In particular, Kapur et al (U.S. Pat. No. 6,268,014, issued October 2000) describe a method for fabricating a solar cell based upon the solution-based deposition of a source material comprised of mechanically milled, oxide-containing, sub-micron sized particles, while Eberspacher and Pauls (U.S. Pat. No. 6,268,014, issued July 2001; U.S. Patent Application Publication 20020006470) describe the forming of mixed metal oxide, sub-micron sized particles by pyrolizing droplets of a solution, then ultrasonically spraying the resulting particles onto a substrate.
However, there are several drawbacks to the use of metal oxides as precursor materials for CIGS solar cells. First, the use of oxide-based particles in CIGS absorber layer construction requires a high-temperature hydrogen reduction step to reduce the oxides. In addition to requiring substantial time and energy, this step is potentially explosive. Further, although it is highly desirable to incorporate gallium in the active layer of the solar cells, the presence of gallium results in the formation of gallium oxide upon annealing, a highly stable material which is very difficult to reduce even under the most extreme conditions. As a result, it is very difficult to effectively incorporate gallium into a nascent copper indium precursor film using a metal oxide synthesis approach.
In addition, the methods of particle formation and deposition taught in the prior art carry significant challenges. For example, mechanical milling is a lengthy process that can requires substantial energy and take several tens of hours to achieve sub-micron sized particles. Further, even after milling, particles are rarely uniform, resulting in a substantial size distribution, which can result in poorly packed precursor films, leading to low-density absorber layers with poor optoelectronic and electronic properties.
Spray pyrolysis of micron thick layers of precursor particles also has significant drawbacks. First, the ultrasonic spraying of thin layers of sub-micron sized particles onto a substrate is an inherently non-uniform process, resulting in differential drying rates as particles are spray deposited. Non-uniform drying can result from any of several factors including but not limited to differential drying on the substrate, mid-stream drying (e.g. drying of droplets prior to reaching a substrate), and pooling of particles and droplets into non-contiguous aggregates that leave space between the aggregation loci. Further, it is especially challenging to achieve any scaling for this technique since it is inherently difficult to carry out a uniform wet deposition of many small particle-containing droplets over a large area without any premature drying prior to completion of the deposition process. Films are often uneven and have substantial spatial non-uniformities across their surface. These and related forms of non-uniform drying lead to the formation of pockets and voids within the deposited film, creating a porous material which leads to a solar cell with poor and unstable optoelectronic and electronic properties. Some of these defects can be overcome when much thicker films are deposited, e.g. in the 20 to 100 micron thick range, but such films are not useful for solar cell devices, which typically require the absorber layer to have a thickness between about one and two microns.
Rapid drying of thin films also limits the potential scaleability of spray deposition techniques. For example, using a scrolling nozzle to spray deposit across a large surface area results in drying of the initially sprayed area prior to deposition of the final area to be deposited. This uneven drying results in additional spatial non-uniformities, pockets, and voids. While the use of multiple spray heads concurrently moving across a substrate surface decreases the time required for deposition across the total surface area, local non-uniformities can arise from regions near each of the nozzles as well as in overlapping regions.
For sub-micron precursor particles formed by either mechanical milling or spray pyrolysis, it is difficult to make precisely-shaped and sized nanoparticles of indium and gallium since indium is so soft and gallium is molten at room temperature. Small and uniformly-sized particles lead to more densely-packed films, which can improve device performance. Furthermore, the properties of the resulting CIGS absorber layer are highly dependent on the stoichiometric ratio of the elements in the layer. The stoichiometric ratios in the layer largely depend on the stoichiometric ratios in the nanoparticles in the precursor solution, while the initial crystalline phase of precursor materials impacts the feasibility of achieving a targeted final phase in the annealed absorber film. It is difficult to precisely tune the stoichiometry and/or phase of the nanoparticles on a nanometer scale with current techniques.
Thus, there is a need in the art, for a non-oxide, nanoparticle based precursor material that overcomes the above disadvantages.